Espectros de RMN como ´ Arboles Binarios

The study showed that the ED length of stay was associated with neither hospital mortality nor HRQoL at six months after intensive care in the 1016 patients admitted to the ICU from the ED (p=.82 and p=.34, respectively). The flowchart for Study I patients is presented in Figure 4.

The median ED length of stay before ICU admission was 4.8 h (280 min, IQR 3.1–7.6 h). Patients admitted to the medical ICU from the ED were younger than patients admitted from elsewhere (median age 56 years vs. 63 years, p=.001), and they had lower disease severity and organ dysfunction scores (median APACHE II score 20 vs. 23, p=.001, median first-day SOFA score 7 vs. 8, p=.001). They also had shorter length of stay in the ICU (ED vs. non-ED patients median 1.56 vs. 2.06 days, p<.001).

Of the 747 patients to whom the 15D questionnaire was sent, 251 (34%) returned it. Five patients had to be excluded from the analysis, as their questionnaires contained more than three missing answers. Responders and nonresponders did not differ from each other in terms of age, gender, length of ED or ICU stay, or SOFA and APACHE II scores (p=.24–.87). The 15D score of the responders (median 0.83, IQR 0.70–0.90) was significantly lower than in the age- and sex-matched general Finnish population, and also on all 15 dimensions (p<.001– p=.008), as presented in Figure 5.

 67   

Intensive care unit n=1537 Emergency

department n=1016 (66%)

High dependency unit or other ICU n=310 (20%) Ward n=144 (9%) Other n=67 (4%) Dead 194 (13%) High dependency unit n=719 (47%) Ward n=433 (28%) Other n=191 (12%)

Figure 4. Patient flow of Study I patients, places of admission, and discharge.

Figure 5. The 15D profiles (mean 15D score) of the Study I patients (n=246) six months after intensive care and the general age- and sex-matched Finnish population.

 68    6.2 Association of cell-free plasma DNA with mortality and severity of disease in

critically ill patients (II)

In Study II, maximum DNA concentration in 228 critically ill patients during the first four days in ICU was higher in hospital nonsurvivors than in survivors (median 9366 GE/ml, IQR 4228–21 159 GE/ml vs. median 6506 GE/ml, IQR 2854–13 745 GE/ml, p=.035; Figure 6). In logistic regression analysis including age, gender, APACHE II score, maximum SOFA score during the first 96 h in the ICU, and maximum plasma DNA; maximum cell-free plasma DNA was independently associated with hospital mortality (p=.049) as well as with maximum SOFA score (p<.001), age (p=.003), and gender (p=.01).

Maximum DNA levels were higher in patients with infection (n=58) than in other patients (median 12 019, IQR 5449–19 583 GE/ml vs. median 5808 GE/ml, IQR 2695–11 645 GE/ml, p<.001). A significant correlation existed between maximum DNA concentration and APACHE II score (r=.17, p=.01) and maximum SOFA score during the first 96 h (r=.25, p<.001), especially with the liver (p=.04), coagulation (p=.02), renal (p=.01), cardiovascular (p=.001), and pulmonary (p<.001) SOFA scores. The maximum cell-free plasma DNA concentration was not associated with age, gender, or mechanical ventilation.

Figure 6. Maximum cell-free plasma DNA concentrations in hospital survivors and nonsurvivors (line: median, box: interquartile range, whiskers: 95% confidence interval).

 69    6.3 Predictive value of plasma DNA in patients with severe sepsis or septic shock (III)

The median cell-free plasma DNA concentration at baseline was 8070 GE/ml (IQR 3883–18 934 GE/ml) and 7457 GE/ml 72 h thereafter (IQR 3668–16 311 GE/ml) in patients with severe sepsis or septic shock. The cell-free plasma DNA concentration was higher in ICU nonsurvivors than in survivors at admission (median 15 904 vs. 7522 GE/ml, p<.001) as well as 72 h later (median 15 176 vs. 6758 GE/ml, p=.004). Hospital nonsurvivors had higher plasma DNA levels than survivors at both time-points (baseline median 12 386 vs. 7678 GE/ml, p=.009 and 72-hour median 11 428 vs. 6414 GE/ml, p=.008). The increasing or decreasing plasma DNA concentration was not associated with ICU or hospital mortality (p=.42–.93).

The AUC regarding ICU mortality was 0.71 (95% CI 0.62–0.80) for the DNA concentration at inclusion and 0.70 (95% CI 0.57–0.82) for the DNA concentration 72 h later (Figure 7). Compared with plasma DNA, the maximum lactate value on the first day as well as the SAPS II score had slightly higher AUCs for ICU mortality (0.77, 95% CI 0.68–0.85, and 0.75, 95% CI 0.65–0.84, respectively), whereas the first-day SOFA score had slightly lower AUC for ICU mortality (0.69, 95% CI 0.58–0.80). Regarding hospital mortality, plasma DNA had less discriminative power; the AUC for the baseline DNA concentration was 0.61 (95% CI 0.53– 0.69) and for the 72-h DNA concentration 0.63 (95% CI 0.53–0.72). In ROC analysis, the best cut-off value of plasma DNA at baseline for ICU mortality was 12 000 GE/ml, with a sensitivity of 67% (95% CI 51–80%), specificity of 67% (95% CI 62–72%), positive likelihood ratio of 2.03 (95% CI 1.49–2.76), and correct classification rate of 67%. The best cut-off value of plasma DNA after 72 h was 12 500 GE/ml, with a sensitivity of 60% (95% CI 39–78%), specificity of 70% (95% CI 64–75%), positive likelihood ratio of 2.00 (95% CI 1.32–3.03), and correct classification rate of 69%.

The 1-year mortality rate in Study III was 40% (102/255). The Kaplan-Meier survival curve for the best plasma DNA cut-off value at baseline is presented in Figure 8.

The Study III patients did not differ from the rest of the Finnsepsis study cohort concerning basic characteristics, severity of disease or organ dysfunction, or mortality.

 70    Figure 7. ROC curves for plasma DNA concentration at baseline (AUC 0.71, 95% CI 0.62–

0.80) and 72 h later (AUC 0.70, 95% CI 0.57–0.82) regarding mortality in the intensive care unit.

Figure 8. Kaplan-Meier survival curve according to plasma DNA concentration at baseline. The cut-off value is the best for predicting ICU mortality.

 71    The variables significantly associated with hospital mortality (SAPS II score minus SAPS age

points, admission SOFA score, age, lactate value, cardiovascular comorbidity, and plasma DNA value at inclusion and 72 h later) were included in multiple logistic regression analysis. The SAPS II score (p=.001) and cardiovascular comorbidity (p=.015) were independent predictors of hospital mortality. The variables significantly associated with ICU mortality (SAPS II score minus SAPS age points, admission SOFA score, age, lactate value, and plasma DNA value at inclusion and 72 h later) were included in separate multiple logistic regression analysis. The first-day plasma DNA concentration (p=.005) and the SAPS II score (p=.008) were independent predictors of ICU mortality.

The variables significantly correlated with baseline cell-free plasma DNA concentration were first-day SOFA score (r=0.29, p<.001), maximum SOFA score (r=0.30, p<.001), APACHE II score (r=0.18, p=.005), SAPS II score (r=0.22, p=.001), maximum first-day lactate concentration (r=0.40, p<.001), first-day plasma creatinine concentration (r=0.15, p=.018), first-day lowest estimated creatinine clearance (r=0.16, p=.013), first-day lowest platelet count (r=0.24, p<.001), first-day lower urine output (r=0.24, p<.001), and first-day urea concentration (r=0.21, p=.005). Neither age nor gender correlated with DNA concentrations (p=.32–.77 and p=.06–.77, respectively). First-day maximum lactate value (p=.003) and SOFA score on the first day of intensive care (p=.015) were independently associated with plasma DNA concentration at baseline in linear regression analysis.